This topic describes the parameters that are used to monitor the performance of compute nodes, data nodes, Global Meta Service (GMS) nodes, and change data capture (CDC) nodes in PolarDB-X instances.

Parameters related to compute nodes

API parameter Description
Cpu_Usage The CPU utilization as a percentage value.
Mem_Usage The usage of heap memory of a Java Virtual Machine (JVM) as a percentage value.

The value of this parameter is calculated based on the following formula: Usage of JVM heap memory = Amount of the used JVM heap memory/Total amount of JVM heap memory × 100%.

Logic_RT The average response time of requests that are sent by users. Unit: microseconds.

The value of this parameter is calculated based on the following formula: Response time of a request = Point in time at which the result of the request is returned - Point in time at which the request is received.

Physical_RT The average response time of requests that are sent from compute nodes to data nodes. Unit: microseconds.

The value of this parameter is calculated based on the following formula: Response time of a request = Point in time at which the compute node receives the result from the data node - Point in time at which the compute node sends the request to the data node.

Logic_QPS The average number of user requests that are processed per second.
Logic_TPS The average number of transactions that are processed per second.
Physical_QPS The average number of requests that are sent to data nodes per second.
Network_Usage The network throughput, which indicates the sum of the inbound traffic and the outbound traffic per second. Unit: byte/s.
Connections The number of connections. The following data can be returned:
  • active_connection: the total number of connections.
  • connection_count: the number of connections that are created per second.
Thread_Detail The number of active connections.
JVM_CollectorTime The average period of time that is required to perform garbage collection (GC) per second. The unit of the value is nanoseconds. The following data can be returned:
  • The average period of time that is required to perform GC in the young generation space per second.
  • The average period of time that is required to perform full GC per second.
Slow_RequestCount

The following data can be returned:

  • The average number of slow SQL requests that are sent by users per second.
  • The average number of slow SQL requests that are sent to data nodes per second.

Parameters related to data nodes or GMS nodes

API parameter Description
MySQL_NetworkTraffic The following data can be returned:
  • The average amount of inbound traffic to the PolarDB-X instance per second. Unit: KB.
  • The average amount of outbound traffic from the PolarDB-X instance per second. Unit: KB.
MySQL_QPSTPS The following data can be returned:
  • The average number of SQL statements that are executed per second.
  • The average number of transactions that are processed per second.
MySQL_Sessions The following data can be returned:
  • The number of active connections.
  • The total number of connections.
MySQL_InnoDBBufferRatio The following data can be returned:
  • The read hit ratio of the InnoDB buffer pool.
  • The usage of the InnoDB buffer pool.
  • The dirty ratio of the InnoDB buffer pool.
MySQL_InnoDBDataReadWriten The following data can be returned:
  • The average volume of data that InnoDB reads per second. Unit: KB.
  • The average volume of data that InnoDB writes per second. Unit: KB.
MySQL_InnoDBLogRequests The following data can be returned:
  • The average number of read operations that are performed on an InnoDB buffer pool per second.
  • The average number of write operations that are performed on an InnoDB buffer pool per second.
MySQL_InnoDBLogWrites The following data can be returned:
  • The average number of log write requests that are processed per second.
  • The average number of physical writes that are performed on log files per second.
  • The average number of operations that are performed by using the fsync() function to write data to log files per second.
MySQL_TempDiskTableCreates The number of temporary tables that are automatically created on the hard disk during statement execution.
MySQL_COMDML The following data can be returned:
  • The average number of times for which the DELETE statements are executed per second.
  • The average number of times for which the INSERT statements are executed per second.
  • The average number of times for which the INSERT_SELECT statements are executed per second.
  • The average number of times for which the REPLACE statements are executed per second.
  • The average number of times for which the REPLACE_SELECT statements are executed per second.
  • The average number of times for which the SELECT statements are executed per second.
  • The average number of times for which the UPDATE statements are executed per second.
MySQL_RowDML The following data can be returned:
  • The average number of rows that are read from an InnoDB table per second.
  • The average number of rows that are updated in an InnoDB table per second.
  • The average number of rows that are deleted from an InnoDB table per second.
  • The average number of rows that are inserted into an InnoDB table per second.
  • The average number of physical writes that are performed on log files per second.
MySQL_MemCpuUsage The following data can be returned:
  • The CPU utilization of the PolarDB-X instance.
  • The memory usage of the PolarDB-X instance.
MySQL_IOPS The IOPS of the PolarDB-X instance.
MySQL_DetailedSpaceUsage The following data can be returned:
  • The used storage of the PolarDB-X instance.
  • The data storage.
  • The log storage.
  • The temporary storage.
  • The system storage.
slavestat The replication delay between the primary instance and the read-only instances.
MySQL_ThreadStatus The following data can be returned: the number of active threads and the number of connected threads.
MySQL_ReplicationDelay The replication delay between the primary instance and the secondary instance. Unit: seconds.
MySQL_ReplicationThread The following data can be returned: the states of I/O threads and SQL threads based on which data can be replicated. The value 1 indicates that the related threads are normal. The value 0 indicates that the related threads are lost.
IOPSUsage The IOPS usage as a percentage value.
MySQL_Slow_Queries The average number of slow SQL requests that are processed per second.

Parameters related to CDC nodes

API parameter Description
Cdc_DumperDelay The latency that occurs when binary log events are being processed. Unit: milliseconds.
Cdc_DumperThroughput The data size of binary log events that are processed per second. Unit: byte/s.
Cdc_CpuUsage The CPU utilization as a percentage value.
Cdc_DumperHeapUsage The usage of JVM heap memory as a percentage value.

The value of this parameter is calculated based on the following formula: Usage of JVM heap memory = Amount of the used JVM heap memory/Total amount of JVM heap memory × 100%.

Cdc_DumperOps The information about binary log events that are processed. The following data can be returned:
  • dumper_total_eps: the average number of binary log events that are processed per second.
  • dumper_dml_eps: the average number of binary log events during which DML statements are executed per second.
  • dumper_tps: the average number of transactions that are processed per second.